Le 27 janvier 2022 05:13:09 GMT+01:00, Richard Stallman <rms@gnu.org> a écrit :
[[[ To any NSA and FBI agents reading my email: please consider ]]]
[[[ whether defending the US Constitution against all enemies, ]]]
[[[ foreign or domestic, requires you to follow Snowden's example. ]]]
> https://www.wired.com/story/22-year-old-builds-chips-parents-garage/
He has made a chip with 1400 transistors. I think that
making a processor capable of running GNU/Linux like the 15-year-old
processor in my T400s will take at least another decade, and probably
two or three.
I'm sure people will get there someday. But RYF's policies should
be designed for the next few years, not for long-term policies.
We can't push long-term policies, we can only encourage them.
--
Dr Richard Stallman (https://stallman.org)
Chief GNUisance of the GNU Project (https://gnu.org)
Founder, Free Software Foundation (https://fsf.org)
Internet Hall-of-Famer (https://internethalloffame.org)
_______________________________________________
libreplanet-discuss mailing list
libreplanet-discuss@libreplanet.org
https://lists.libreplanet.org/mailman/listinfo/libreplanet-discuss
Hello to all. I add my little stone to this discussion to say first that I
agree with Richard that today we cannot yet say no to non-free hardware as we
do with software, and that we cannot push such pre-requisites into RYF.
What I have to add is mostly on the subject of processors, since we are talking
about someone who has designed his own CPU.
Some of you know that I'm a student in France in a master's degree in processor
architecture. My laboratory, the LIP6, has been trying for a few years to focus
on VLSI (Very Large Scale Integration) design methods to design free hardware.
Thus, our university has created a complete VLSI chain as free software. It is
called Coriolis and allows the design of an ASIC from its VHDL description to
the drawing of the masks (see https://coriolis.lip6.fr)
Moreover, to design elements more quickly, component libraries are needed (to
avoid reinventing the wheel). Especially for analog circuits. Our laboratory
has therefore created the OCEANE library (sorry link in French, nobody has
translated it yet https://www-soc.lip6.fr/equipe-cian/logiciels/oceane/).
Finally, so that the masks produced by these methods can be transformed into
real hardware, we had to find a foundry that would give us its PDK (all the
information to correctly size the masks and adapt them to specific
technologies). Unfortunately most foundries keep these parameters a secret. But
recently, SkyWater foundry released their PDK. It is a technology with a fine
130nm etch. It's not state of the art, but it's enough to do something useful.
You can find it here:
https://skywater-pdk.readthedocs.io/en/main/
Thanks to this, at LIP6 (which belongs to Sorbonne University) we are now able
to implement processors respecting the ARM v2 specification freely. A work on
RISC V is in progress this year.
OpenPGP_0x8EA91B51E27E3D99.asc